Selected Frailty Testing Can Predict for Risk of Febrile Neutropenia in Multiple Myeloma Patients Undergoing Autologous Stem Cell Transplant

Introduction:

Validated myeloma-specific frailty assessment tools, such as the IMWG geriatric assessment tool (IMW GAT) and the revised myeloma co-morbidity index (R-MCI), can be useful to tailor treatment intensity in MM patients (pts). To date, most MM studies of frailty have focused on older, transplant-ineligible populations, with less known about the utility of frailty assessments to guide therapeutic decision-making in younger, transplant pts. We previously reported that frailty scores largely validated in older MM cohorts correlate well with each other, but generally lack ability to discriminate between unfit and fit patients, when applied to a younger transplant population (Devasia et al ASH 2022). Here, we evaluate the use of pre-transplant frailty assessment, using a battery of subjective and objective tools, for predicting common post-transplant toxicities, such as febrile neutropenia (FN) and organ complications.

Methods:

162 consecutive MM pts preparing for ASCT and consented for an ongoing prospective frailty study were included. Prior to ASCT, patients underwent comprehensive frailty assessments including: 1) objective measures of fitness [hand grip strength, 6-min walk test (6MWT), Timed Up and Go (TUG), 2) subjective measures of function (Fried Frailty Phenotype (FFP), Rockwood Frailty, KPS, ECOG, HSCT-CI, ESAS), 3) scoring systems integrating both objective and subjective measures (R-MCI, IMWG geriatric assessment tool (IMWG-GAT), and 4) routine screening for organ function (CrCl, serum albumin, hemoglobin, BNP, PFT, and cardiac imaging). Following ASCT, acute grade 3-4 organ-specific toxicities (including FN) occurring from stem cell infusion through engraftment to discharge were prospectively collected (NCI-CTCAE criteria V5.0). Patient/disease characteristics, stem cell dose, engraftment, and transfusion requirements were also collected from electronic patient records. Chi-square and/or Fisher's exact test, as appropriate, were used to assess the association of categorical covariates with frailty scores/measures. Logistic regression was used to assess the association of potential predictors to binary outcomes. All p-values were 2-sided and p<0.05 was considered statistically significant.

Results:

In total, 162 pts were included in this analysis, median age 62 yrs (range 35-73). Most pts received melphalan 200 mg/m2 conditioning, with 10 pts (6%) dose-reduced to 140mg/m2. All patients received antibiotic prophylaxis. Median stem cell dose was 2.82 x 10 ⁶ CD34+ cells/kg Median neutrophil and platelet engraftment times were 13(10-17) and 14(8-45) days, respectively.

104(64%) pts developed FN post-transplant. Other grade 3-4 toxicities included: GI (41 pts; 25.3%), cardiac (12 pts; 7.4%), lung (4 pts; 2.4%). When correlating with pre-transplant frailty assessments, ECOG performance status and R-MCI appeared to predict for risk of FN, but not for specific organ toxicity (GI, cardiac) nor presence of >1 grade 3-4 toxicity (Table 1). In pts with ECOG 1-2 at baseline, the incidence of FN was 70.9% (83/117) compared to ECOG 0 pts 46.7% (21/45)(p=0.0039). There were no patients with ECOG 3-4 on study. Similarly, 58% (59/101) with baseline low risk frailty as per R-MCI developed FN, compared to 73.8% (45/61) in either R-MCI intermediate or high risk groups (p=0.050).

We then evaluated the combined effect of ECOG and R-MCI on prediction of FN. Patients with ECOG 0 plus R-MCI low risk were grouped as “Low risk FN”, patients with ECOG 1-2 plus R-MCI intermediate/high risk were grouped as “High risk FN”. All other patients were grouped as “Intermediate risk FN” (Table 2). By using 2 scores, we could better predict for FN, with 72.5% in “High risk FN” compared with only 37% in “Low risk FN” (p=0.0008)].

Conclusions:

This preliminary analysis of an ongoing frailty study in MM patients undergoing ASCT suggests potential utility of combining ECOG and R-MCI in predicting for development of FN. In pts with both ECOG 0 and R-MCI low risk (21% of total cohort), FN occurred in only 37%. Although further validation in a larger cohort is required, this finding may help to select at-risk patients for prophylactic peri-transplant antibiotics whilst avoiding routine use in all patients. This has clinical significance, given the emerging recognition of potentially detrimental effect on the gut microbiota and risk of antibiotic resistance in this setting.